WHITE PAPER

BENEFITS OF MIGRATING IVR SERVICES TO SIP/RTP AS A PRELUDE TO NEW MEDIA SERVICE DELIVERY

CONTENTS EXECUTIVE SUMMARY MEDIA SERVICES ARCHITECTURE THE DISTRIBUTED MEDIA ADVANTAGE WHY NOT DISTRIBUTE THE SERVICE CONTROL? NGN, IMS AND IVR TERMINOLOGY CO-EXISTENCE AND MIGRATION INCENTIVES TO MIGRATE TO SIP/RTP MORE PERSPECTIVE ON SAVINGS WITH RING BACK TONES POWER AND FLEXIBILITY IN A SERVICE PLATFORM SERVICE DELIVERY ECO-SYSTEM CONCLUSION ABOUT THE AUTHOR 1 2 2 3 3 4 4 4 5 5 6 6

EXECUTIVE SUMMARY

All communications service providers including mobile operators are making increasing use of packet-switched technology such as an MPLS core, soft switch, and soft IVR as they transition to NGN and IMS architectures. In particular, the use of a soft IVR means that many existing media services can be migrated to more efficient SIP/RTP technology to produce costs savings and more flexible solutions. Media based services, such as self-service IVR systems, Ring Back Tone servers, service announcements, or call centres, are resource intensive and require highly scalable architectures. By adopting a distributed architecture and a platform that combines modules for handling both PSTN-based and IP-based resources an orderly migration to a full NGN solution can be achieved with no disruption to the customer base and negligible business risk. And of course media solutions that are pre-integrated with other service delivery platform capabilities produce further efficiencies as well as a compelling opportunity for next generation service innovation.

01
PAGE

MEDIA SERVICES ARCHITECTURE

A common feature of both PSTN and NGN communications networks is the separation of transmission facilities into signalling and bearer networks. This allows a generalization of the IVR or media application architecture into central control functions and distributed media functions.
SERVICE CREATION SERVICE CREATION CALL CONTROL MEDIA CONTROL

THE DISTRIBUTED MEDIA ADVANTAGE

Each Voice circuit in PSTN requires a 64Kbps channel over T1/E1 links and requires committed resources for any call. Hence placing media servers as close as possible to the ingress point, such as the MSC handling the call, minimizes the use of voice trunks between switching centres. If all media servers were centralized then every IVR call would be carried to the primary site in order to Play Announcements and Collect User Information (the PACUI commands shown above) or play a Ring Back Tone.

SERVICE EXECUTION CONTROL FUNCTIONS

DISTRIBUTED FUNCTIONS

MSC or SOFT SWITCH

MEDIA SERVER

Example 1: If a remote switching site supports a population of 500K subscribers with on average 0.1 Busy Hour Call Attempts (BHCA) per subscriber to the IVR, and an average call hold time of 30 seconds, then we need: 500,000 x 0.1 x 30 / 3,600 = 417 voice circuits

PSIN

These circuits must be dedicated between the switch and the media server location for IVR activities. Of course this also means that 417 IVR ports are required for this service.
FIGURE 2 MULTI-SITE DISTRIBUTED MEDIA FOR AN SS7 NETWORK

Using componentized software principles it is possible to distribute functional modules to any location so an optimal distribution takes a number of factors into account including type of traffic handled, integration with other systems, and ease of management.

Placing IVR port capacity adjacent to the serving MSC saves 417 inter-switch voice trunks, or 14 x E1 links. And high media usage, e.g. popular Ring Back Tone services, means even higher savings using such a distributed media plane.

FIGURE 2 MULTI-SITE DISTRIBUTED MEDIA FOR AN SS7 NETWORK

PRIMARY SWITCHING SITE

BSS APPLICATION

OTHER APPLICATION

APPLICATION GATEWAY

SYSTEMS INTEGRATION

SERVICE CREATION

SERVICE CONTROL

SS7/ISUR CALL CONTROL

SS7/CS1 or IP-BASED PACUI INSTRUCTIONS

MEDIA FILE MANAGEMENT

MEDIA SERVER

SS7 CORE

MEDIA SERVER

MSC

MSC

SWITCH SITE A

SS7 SIGNALING E1/VOICE TRUNKS

SWITCH SITE B

02
PAGE

WHY NOT DISTRIBUTE THE SERVICE CONTROL?

Having justified a distributed media plane so that it avoids the carriage of IVR media traffic to a central site the question arises as to why we should not treat the signalling plane in the same way and distribute the Service Control Function (SCF). It can be noted that whilst a distributed SCF may have benefits in certain scenarios, there are a number of major trade-offs that must be considered in relation to media service delivery. > Signalling traffic requires significantly less bandwidth than media traffic > Signalling is likely to be centrally managed anyway for other reasons, such as prepaid charging > Most IVR services require access to subscriber data or other centrally managed applications, so some type of transaction traffic must be directed to the primary site to complete the IVR service > The advantages in using SS7 for routing, load-sharing and failure handling across large geographies are well proven as part of the core network > The point-to-point integration (or even EAI bus) used for IVR application integration is not standardized or conducive to multi-site routing and operational management > Since the primary site is likely to be the seat of management, including service development and testing, a central SCF simplifies operations and hence reduces the overall solution costs There are many variables involved when it comes to signalling traffic calculations so no precise figures can be provided here. However, it is should be noted that the SS7/ISUP traffic between the MSC and SCF for IVR call handling is relatively simple and lightweight, as is the PACUI traffic from SCF to media servers. In addition, all signalling traffic can share point-to-point SS7 links. IVR applications vary, and traffic from the SCF to back-end subscriber databases or integrated applications depends on the IVR service. It is reasonable to conclude that any traffic/bandwidth savings that might arise by distributing SCF do not compensate for the added complexity, costs, and risks associated with a distributed SCF architecture. Unless there are other reasons not related to media delivery the suggested architecture in Figure 2 is highly recommended for optimal media delivery outcomes, including multi-regional IVR solutions for mobile operators. Inter-site bandwidth can easily demonstrate 90% savings (and often much more) for a distributed IVR service.

NGN, IMS AND IVR TERMINOLOGY

Now we turn our attention to the migration of IVR services to NGN and discover that this transition to IP networking does not change the distributed architecture principles. For NGN we substitute SIP for SS7/ISUP in the signalling plane, and RTP sessions for E1/Voice bearers across the media plane. In functional terms, the Specialized Resource Function (SRF) or Intelligent Peripheral (IP) that provided basic IVR facilities for the SSP or SCP in the IN architecture is replaced in IMS by the Media Resource Function (MRF). Similarly, the SCF becomes the SIP Application Server (SIP/AS). In fact, IMS further divides the MRF into the Media Resource Function Controller (MRFC) in the signalling plane and the Media Resource Function Processor (MRFP) in the media plane then

FIGURE 3 MULTI-SITE DISTRIBUTED MEDIA FOR NGN IMS

SERVICE CREATION

PRIMARY SWITCHING SITE

BSS APPLICATION

OTHER APPLICATION

APPLICATION GATEWAY

SYSTEMS INTEGRATION

MRFC

SERVICE CONTROL

SIP AS

SIP/SDP CALL CONTROL

IP-BASED (H.248) PACUI INSTRUCTIONS

MRFP MEDIA SERVER

IP CORE

MRFP MEDIA SERVER

CSCF

SOFT SWITCH

CSCF

SOFT SWITCH

SIP SIGNALING RTP SESSIONS PSTN SWITCH SITE A SWITCH SITE B PSTN

03
EXISITNG IVR COMPONENTS SUPPORTING COMPONENTS IMS COMPONENTS

PAGE

stipulates the use of H.248 as the control protocol between these in an effort to define standard components and interfaces. For practical reasons the media server capabilities, including both MRFC and MRFP, are almost always provided by a single vendor and the use of H.248 is not widely accepted. Indeed most vendors have simply evolved their existing IVR architecture into IP-based multimedia solutions.

In other words, RTP media servers can offer a 70% saving on bandwidth between the switch and the media servers. The bandwidth saving comes on top of the savings achieved by media plane distribution, and the generally lower costs of an IP core over the traditional SS7/E1 core. Finally, there is one additional area of cost saving that arises in solutions such as those offered by Amdocs. When using SIP/RTP media servers a higher port density (meaning more concurrent calls) can be handled on a given hardware server than if the same platform was serving identical media services using SS7/E1 technology. All of this adds up to a much improved price/performance to deliver on the benefits of NGN.

CO-EXISTENCE AND MIGRATION

The first step in a full migration to an NGN architecture is coexistence. This may take one of 2 main forms: > Side-by-side deployments > Hybrid (dual support) components Supporting side-by-side deployments of similar components may prove costly if the migration takes longer than expected, especially if different vendors are involved. As well as any functional duplication, the management of different systems may add complexity and increase the cost of service maintenance. Hybrid solutions like those from Amdocs that support IVR and media services over either SS7 or IP networks leverage a layered architecture where multiple protocols can be comfortably handled together under a common service execution platform. Using this approach the migration can proceed in situ from the existing IVR services to new IP-based soft IVR services at a pace based solely on business and operational readiness.

MORE PERSPECTIVE ON SAVINGS WITH RING BACK TONES

The good news on price/performance advantages gets even better in the face of large scale media solutions. Consider some popular services, such as the Personal Ring Back Tone (or Caller Tone) service. This is a media resource intensive service and can have extremely high scalability. A subscriber using this has all incoming calls connected to a media server to play an alternate caller tone, such as a favourite song or jingle. Here is how the savings with SIP/RTP media are magnified on such a service.

INCENTIVES TO MIGRATE TO SIP/RTP

The lure of NGN solutions is lower costs. Up front costs are reduced due to the commoditization and standardization of components, but most savings come about through the lower transmission costs involved when using IP-based solutions to deliver services to the end user. Example 3: Taking the switching site population of 500K subscribers used in Example 1, we will assume that 40% of subscribers register for the Caller Tone service, and generally there is 1 incoming call per subscriber during busy hour. If the average network ring time is 15 seconds then we can calculate the number of media ports as follows: 500,000 x 40% x 1 x 15 / 3,600 = 833 ports Example 2: Take a typical IVR slice of 240 ports, meaning we need to be able to concurrently handle 240 user sessions for announcements and/or Dual Tone Multi-Frequency (DTMF) user interaction, like pressing menu options. The 240 port bandwidth required using E1/Voice is about 16Mbps whilst for RTP using G729A or G726 it is about 4.8Mbps.

This means that whereas for a standard IVR service the mobile operator may require approximately 1 media port for every 1,000 subscribers, when using Caller Tone services the ratio may be closer to 1 media port for every 500 subscribers, or even less. Hence the savings realized through higher port density, lower media bandwidth requirements, and IP-based network infrastructure, are magnified with Caller Tone services. With a future that promises more media-based services for the end user as part of the shift to a more open, Internet style, services regime underpinned by the IMS architecture, the time to think about NGN migration is now.

04
PAGE

POWER AND FLEXIBILITY IN A SERVICE PLATFORM

So far the focus in this paper has been on the technical merits of migrating media based services, such as a standard IVR found in every mobile network, to a solution using IP-based services, such as those specified by IMS. It has been demonstrated that such a transition offers significant cost savings. It has also been shown that by adopting proven architectures and using component based solutions such a transition can be made simple and low risk. But when we consider service delivery with multimedia services, we must also consider how media resources can be reused as building blocks to construct more sophisticated and marketable services for subscribers, and how various services, both within and external to the operators environment, can be aggregated and used in innovative ways to create competitive value added services. We start with a closer look at the Amdocs system core.
GUI DESIGNER APPLICATION SERVER

The core technology involves 2 main components: > Service Creation that provides a GUI design, debug, and deployment environment for any service and can use over 300 Service Independent Building Blocks (SIBs) inbuilt, pre-tested service functions > Application Server which runs the designed script on a extremely scalable, high performance engine, with access to the service delivery eco-system A service instance executes a script that is invoked by an external trigger such as a network transaction. The invoke request is a set of Attribute Value Pairs (AVP) so that the engine is entirely independent of any protocol. This means that it easy to support common services even across many protocols. Each script may in turn invoke other onplatform or off-platform applications to enrich the service logic.

SERVICE DELIVERY ECO-SYSTEM

SERVICE CONTROL SCRIPT LOGIC SERVICE EXECUTION

SERVICE CREATION

A full discussion on service delivery and control is beyond the scope of this paper but by now we should recognize that an IVR, or Ring Back Tone server, are just examples of media-based applications. As the old stove-piped view of services gives way to the service delivery eco-system based on horizontal layers, component architecture, standard interfaces, and open access, we can now position media services in the context of end-to-end service delivery. Media service delivery becomes part of the overall service capability mix, leveraging the Amdocs core system technology for flexible service design and execution, and integrating with other eco-system elements such as Online Charging Systems (OCS) for access to subscriber account details, Diameter for charging, and SOAP to invoke Web Services or applications within, or external to, the operator.

INVOKE

INVOKE

FIGURE 4 GENERAL PURPOSE SERVICE DESIGN AND EXECUTION

SERVICE CONTROL CONTENT MANAGEMENT SOAP SERVICE DATA SERVICE ORCHESTRATION SERVICE CREATION CALL CONTROL MEDIA CONTROL DIAMETER SUBSCRIBER AND CHARGING EXTERNAL APPLICATIONS

SERVICE EXECUTION SERVICE TRIGGER PROTOCOL CONTROL CAP WIN MAP ISUP SIP XML INAP SMPP ETC MEDIA CONTROL

MSC or SOFT SWITCH

MEDIA SERVER

MSC or SOFT SWITCH

MEDIA SERVER

FIGURE 5 MEDIA-BASED SERVICE DELIVERY ECO-SYSTEM

05
PAGE

CONCLUSION
Benefits gained by shifting to NGN technology for media services such as IVR and RBT are then inherited by any subsequent media applications. New marketing initiatives can build on the strength of integrated Call Control and Media Control within this eco-system to create service variants that stimulate customer loyalty and open the way for new revenues. Examples of service evolution that are enabled include: > Video Ring Back which evolves the popular Caller Tones (or RBT) service into a full multimedia play > Interactive Voice and Video (IVV) which adds colour to familiar IVR services by adding video to audio instructions and speech as well as DTMF interaction > Mobile Blogging where users can share media as a way of social networking > Mobile Advertising to insert payed advertisements into subscriber Call or Message services which will widen revenue opportunities and offset declining usage charges > Targeted offers using Business Intelligence and market segmentation tools to target individual subscribers for up-sell and cross-sell > Mobile Shopping to deliver an online shopping experience to time challenged subscribers building loyalty and value Such next generation services, and many others like these, expect to co-exist in a vibrant eco-system of multimedia interaction. The foundation is a next generation services platform based on a proven future-proof architecture. Amdocs can support the present needs of mobile operators for services as common as IVR and RBT on the Amdocs Service Platform. However, by using this platform these services and many other integrated subscriber transactions can be seamlessly migrated into the NGN world. The core Service Creation and Application Server technology provides a single platform that equally supports all protocols and orchestrates service delivery through open integration with resources in other systems. Multimedia based services are an inherent part of this service delivery eco-system making the promised benefits of the next generation of services in reach today

ABOUT THE AUTHOR

Paul Wagner is a Solutions Director at the Amdocs Network Business Unit and a communications consultant with over 25 years experience in system architecture, distributed transaction processing, large scale solution design, service delivery systems, and telecommunications standards. This background yields practical business and technical insights into the ongoing challenges of convergence across the Information Technology and Telecommunications industries. paul.wagner@amdocs.com

09
PAGE

EXECUTIVE SUMMARY

ABOUT AMDOCS NETWORK BUSINESS UNIT

The Amdocs Network Business Unit offers network-connected service delivery and convergent charging solutions designed to address the requirements and cost expectations of service providers in high-growth markets. These solutions provide unmatched operational flexibility and fast time-to-market, enabling service providers to rapidly and cost-effectively create, deliver, manage, charge and profit from innovative new value-added-services (VAS), across any network and any line-of-business for true competitive advantage and growth. For more information, visit http://www.amdocs-network.com

ABOUT AMDOCS

Amdocs is the market leader in customer experience systems innovation, enabling world-leading service providers to deliver an integrated, innovative and the intentional customer experienceat every point of service. Amdocs provides solutions that deliver customer experience excellence, combining the software, service and expertise to help its customers execute their strategies and achieve service, operational and financial excellence. A global company with revenue of $2.84 billion in fiscal 2007, Amdocs has over 16,000 employees and serves customers in more than 50 countries around the world. For more information, visit www.amdocs.com.

Amdocs has offices, development and support centers worldwide, including sites in:
THE AMERICAS: BRAZIL CANADA MEXICO UNITED STATES ASIA PACIFIC: AUSTRALIA CHINA INDIA JAPAN THAILAND EUROPE, MIDDLE CYPRUS CZECH REPUBLIC FRANCE GERMANY HUNGARY EAST & AFRICA: IRELAND ISRAEL ITALY NETHERLANDS POLAND RUSSIA SOUTH AFRICA SPAIN SWEDEN TURKEY UNITED KINGDOM

For the most up-to-date contact information for all Amdocs offices worldwide, please visit our website at www.amdocs.com/corporate.asp

Copyright Amdocs 2008. All Rights Reserved. Reproduction or distribution other than for intended purposes is prohibited, without the prior written consent of Amdocs. Amdocs reserves the right to revise this document and to make changes in the content from time to time without notice. Amdocs may make improvements and/or changes to the product(s) and/or programs described in this document any time. The trademarks and service marks of Amdocs, including the Amdocs mark and logo, Ensemble, Enabler, Clarify, Return on Relationship, DDP/SQL, DDP/F, Intelecable, STMS, Collabrent and intentional customer experience are the exclusive property of Amdocs, and may not be used without permission. All other marks are the property of their respective owners. WP/BNFTS_IVR_SRVCS_05.08